Matins
← All briefs
Matins
Wednesday 22 April — v2.1.117
28 changes / 5 actionable / 3 deep dives
Claude Code

  • Default effort is now high for Pro/Max subscribers on Opus 4.6 and Sonnet 4.6. If you preferred the old medium default, run /effort medium to restore it (more below).

  • Native builds on macOS and Linux replace Glob and Grep tools with embedded bfs and ugrep via Bash. Searches are faster but go through Bash instead of separate tool calls; npm and Windows builds are unchanged (more below).

  • Opus 4.7 sessions were autocompacting far too early. Claude Code was computing context against a 200K window instead of Opus 4.7's native 1M (more below).

  • /model selections now persist across restarts, even when the project pins a different model. The startup header also shows when the active model comes from a project or managed-settings pin.

  • cleanupPeriodDays now sweeps ~/.claude/tasks/, shell-snapshots/, and backups/. If you rely on old task files or backups surviving cleanup, check your retention setting.

New in 2.1.117

2.1.117 (April 22, 2026)

  • Forked subagents can now be enabled on external builds by setting CLAUDE_CODE_FORK_SUBAGENT=1
  • Agent frontmatter mcpServers are now loaded for main-thread agent sessions via --agent
  • Improved /model: selections now persist across restarts even when the project pins a different model, and the startup header shows when the active model comes from a project or managed-settings pin
  • The /resume command now offers to summarize stale, large sessions before re-reading them, matching the existing --resume behavior
  • Faster startup when both local and claude.ai MCP servers are configured (concurrent connect now default)
  • plugin install on an already-installed plugin now installs any missing dependencies instead of stopping at "already installed"
  • Plugin dependency errors now say "not installed" with an install hint, and claude plugin marketplace add now auto-resolves missing dependencies from configured marketplaces
  • Managed-settings blockedMarketplaces and strictKnownMarketplaces are now enforced on plugin install, update, refresh, and autoupdate
  • Advisor Tool (experimental): dialog now carries an "experimental" label, learn-more link, and startup notification when enabled; sessions no longer get stuck with "Advisor tool result content could not be processed" errors on every prompt and /compact
  • The cleanupPeriodDays retention sweep now also covers ~/.claude/tasks/, ~/.claude/shell-snapshots/, and ~/.claude/backups/
  • OpenTelemetry: user_prompt events now include command_name and command_source for slash commands; cost.usage, token.usage, api_request, and api_error now include an effort attribute when the model supports effort levels. Custom/MCP command names are redacted unless OTEL_LOG_TOOL_DETAILS=1 is set
  • Native builds on macOS and Linux: the Glob and Grep tools are replaced by embedded bfs and ugrep available through the Bash tool, giving faster searches without a separate tool round-trip (Windows and npm-installed builds unchanged)
  • Windows: cached where.exe executable lookups per process for faster subprocess launches
  • Default effort for Pro/Max subscribers on Opus 4.6 and Sonnet 4.6 is now high (was medium)
  • Fixed Plain-CLI OAuth sessions dying with "Please run /login" when the access token expires mid-session. The token is now refreshed reactively on 401
  • Fixed WebFetch hanging on very large HTML pages by truncating input before HTML-to-markdown conversion
  • Fixed a crash when a proxy returns HTTP 204 No Content. Now surfaces a clear error instead of a TypeError
  • Fixed /login having no effect when launched with CLAUDE_CODE_OAUTH_TOKEN env var and that token expires
  • Fixed prompt-input undo (Ctrl+_) doing nothing immediately after typing, and skipping a state on each undo step
  • Fixed NO_PROXY not being respected for remote API requests when running under Bun
  • Fixed rare spurious escape/return triggers when key names arrive as coalesced text over slow connections
  • Fixed SDK reload_plugins reconnecting all user MCP servers serially
  • Fixed Bedrock application-inference-profile requests failing with 400 when backed by Opus 4.7 with thinking disabled
  • Fixed MCP elicitation/create requests auto-cancelling in print/SDK mode when the server finishes connecting mid-turn
  • Fixed subagents running a different model than the main agent incorrectly flagging file reads with a malware warning
  • Fixed idle re-render loop when background tasks are present, reducing memory growth on Linux
  • [VSCode] Fixed "Manage Plugins" panel breaking when multiple large marketplaces are configured
  • Fixed Opus 4.7 sessions showing inflated /context percentages and autocompacting too early. Claude Code was computing against a 200K context window instead of Opus 4.7's native 1M

Notes

Default effort completes the rollout to all user tiers. 2.1.94 flipped the default from medium to high for API-key, Bedrock, Vertex, Foundry, Team, and Enterprise users. 2.1.117 extends that to Pro and Max subscribers. Every subscriber tier now defaults to high. If you notice longer response times or higher token usage and preferred the old behavior, /effort medium restores it.

Native builds are absorbing standalone tools into the Bash layer. 2.1.113 moved the CLI to a native binary. 2.1.117 follows through by replacing the JavaScript Glob and Grep tool implementations with bfs (a breadth-first find) and ugrep embedded in the native binary. The practical effect: searches are faster, but agent code or custom hooks that call Glob(...) or Grep(...) directly will get tool-not-found errors on native builds. If you are on npm install, nothing changes.

Opus 4.7's 1M context window was being wasted. Since Opus 4.7 shipped in 2.1.111, context calculations assumed a 200K window. This meant /context showed inflated usage percentages and autocompaction triggered roughly five times too early, effectively capping usable context at 200K. If you noticed Opus 4.7 sessions compacting much sooner than expected, this is the fix.